Nuclear translocation of cytosolic phospholipase A2 is induced by ATP depletion.

Phospholipase A(2) (PLA(2)) enzymes may play a role in cellular injury due to ATP depletion. Renal Madin-Darby canine kidney cells were subjected to ATP depletion to assess the effects of cellular energy metabolism on cytosolic PLA(2) (cPLA(2)) regulation. ATP depletion results in a decrease in soluble cPLA(2) activity and an increase in membrane-associated activity, which is reversed upon restoration of ATP levels by addition of dextrose. In ATP-depleted cells cPLA(2) mass shifts from cytosol to nuclear fractions. GFP-cPLA(2) is localized at the nuclear membrane of stably transfected ATP-depleted LLC-PK(1) cells under conditions where [Ca(2+)](i) is known to increase. cPLA(2) translocation does not occur if the increase in [Ca(2+)](i) increase is inhibited. If [Ca(2+)](i) is allowed to increase when ATP is depleted and the cells are then lysed, cPLA(2) remains associated with nuclear fractions even if the homogenate [Ca(2+)] is markedly reduced. In contrast, cPLA(2), which becomes associated with the nucleus when [Ca(2+)](i) is increased using ionophore, readily dissociates from the nuclear fractions of ATP-replete cells upon reduction of homogenate [Ca(2+)]. Okadaic acid inhibits the ATP depletion-induced association of cPLA(2) with nuclear fractions. Thus energy deprivation results in [Ca(2+)]-induced nuclear translocation, which is partially prevented by a phosphatase inhibitor.

Phospholipase A 2 (PLA 2 ) 1 enzymes release arachidonic acid from the sn-2 position of membrane phospholipids leaving behind a lysophospholipid. PLA 2 enzymes are activated with ischemia/reperfusion in vivo and ATP depletion of cultured cells in vitro (1)(2)(3). Free fatty acids, lysophospholipids, and the metabolites of arachidonic acid have been implicated in cell and tissue injury associated with ischemia and hypoxia (4 -6).
The group IVA cytosolic PLA 2 (cPLA 2 ) is a large molecular mass member of the family of PLA 2 enzymes. cPLA 2 has selectivity for diacylphospholipids containing arachidonic acid at the sn-2 position, and its activity is regulated over the range that intracellular (cytosolic) free [Ca 2ϩ ] ([Ca 2ϩ ] i ) is regulated (7). cPLA 2 has been shown to be critical for the production of eicosanoids (8,9) and has been associated with cell injury in vitro (10,11) and in vivo (4,8). cPLA 2 Ϫ/Ϫ mice are protected against ischemic (8) and toxic (12) injury to the brain.
The mechanisms regulating the activity of cPLA 2 have not been fully elucidated. Amino acid residues essential for catalysis include Asp-549, Asp-200, and Ser-228 (13). Phosphorylation of serine residues has been associated with an increase in cPLA 2 activity and retardation of electrophoretic mobility (14 -18). Serine 505 has been shown to be a critical phosphorylation site for agonist-induced increase in cPLA 2 activity and arachidonic acid release (15, 19 -21). By contrast, Ser-505 has been shown to be unnecessary for arachidonic acid release from thrombin-stimulated platelets, suggesting that other regulatory phosphorylation sites may play an equally important role (22). Ser-727, Ser-437, and Ser-454 may also be phosphorylated and modulate cPLA 2 activity (23).
Like many, but not all, isoforms of PLA 2 , cPLA 2 is calcium-dependent. Unlike the secretory PLA 2 enzymes, however, cPLA 2 is activated at submicromolar rather than millimolar [Ca 2ϩ ] (7,24). cPLA 2 translocates from the cytosol to membranes, particularly the nuclear membrane and endoplasmic reticulum (25,26), in response to increases in [Ca 2ϩ ] i . The distribution of cPLA 2 activity is determined by the [Ca 2ϩ ] of the homogenate from which membrane and cytosolic fractions are obtained (7,(27)(28)(29). If the homogenization buffer contains a [Ca 2ϩ ] similar to that seen in unstimulated cells, the majority of PLA 2 activity is found in the cytosolic fraction. If the homogenate [Ca 2ϩ ] is increased to levels found in stimulated cells, the majority of activity is demonstrated in the membrane fraction. The mechanism for [Ca 2ϩ ]-dependent translocation is entirely contained within an amino-terminal 134-amino acid residue fragment, which is homologous to C2 domains found in phospholipase C and protein kinase C (PKC) (7,30,31). Functional and structural analyses suggest that cPLA 2 contains two independently folded domains (7,30). The C2 domain is an anti-parallel ␤-sandwich that forms three loops where two [Ca 2ϩ ] ions bind (32). The mutation of any of five putative [Ca 2ϩ ]-binding residues results in a greater [Ca 2ϩ ] requirement for both binding of cPLA 2 to phosphatidylcholine substrate and for cPLA 2 activity (33). Ca 2ϩ enables cPLA 2 to associate with its membrane substrate rather than for catalysis.
A critical duration of [Ca 2ϩ ] elevation is required for persistent membrane localization and full cPLA 2 activation.
Whereas a brief increase in [Ca 2ϩ ] i results in translocation of cPLA 2 without an increase in arachidonic acid (AA), a [Ca 2ϩ ] increase of longer duration results in prolonged translocation and AA release even after [Ca 2ϩ ] has returned to basal levels (34). It is not clear whether an increase in [Ca 2ϩ ] i is the only requirement for cPLA 2 translocation to and association with the membrane, because mutations of hydrophobic residues in calcium-binding regions inhibit the association of cPLA 2 with the membrane, despite the intact ability of the mutated cPLA 2 to bind calcium (31). In addition, agents that do not cause an increase in [Ca 2ϩ ] have recently been shown to increase AA release in wild type mouse peritoneal macrophages but not those derived from cPLA 2 knock-out mice, suggesting that regulatory mechanisms in addition to, or in place of, [Ca 2ϩ ] may play a role in cPLA 2 translocation (21,35).
The exposure of Madin-Darby canine kidney (MDCK) cells to 5 mM cyanide and 5 mM 2-deoxyglucose, in the absence of metabolic substrates, results in a reduction of cellular ATP content to less than 5% of control values within 5 min and a striking increase in AA release (2). We have studied the response of cPLA 2 to ATP depletion in MDCK cells using this established model of chemical anoxia. We have found that ATP depletion results in a shift in PLA 2 activity from soluble to insoluble fractions of cell lysates. cPLA 2 translocates to nuclei of ATP-depleted cells as demonstrated by both Western blot analysis and immunofluorescence. This shift in enzyme mass and activity is partially blocked with okadaic acid, a phosphatase 2A inhibitor, and persists despite a reduction in [Ca 2ϩ ] with 2 mM EGTA, indicating that mechanisms other than an increase in [Ca 2ϩ ] alone are responsible for the translocation of cPLA 2 .

MATERIALS AND METHODS
Cell Culture-MDCK cells (no. CCL 34; American Type Culture Collection, Rockville, MD) were grown in Dulbecco's modified Eagle's medium (Mediatech, Inc., Herndon, VA) with 10% fetal calf serum. Cells were plated in 10-cm dishes and used at confluence.
Induction of Chemical Anoxia-Chemical anoxia was induced by incubating cell monolayers for various time points with 5 mM cyanide and 5 mM 2-deoxyglucose, in the absence of glucose. Cell monolayers were incubated in either a Krebs-Henseleit buffer (KHB) containing (in mM) 115 NaCl, 3.6 KCl, 1.3 KH 2 PO 4 , 25 NaHCO or HEPES-based buffer containing (in mM) 130 NaCl, 3.5 KCl, 1.5 KH 2 PO 4 , 1 CaCl 2 , 1 MgCl 2 , 20 HEPES, pH 7.4. KHB buffer was used for experiments in which the S100 and insoluble fractions were isolated. HEPES buffer was used for experiments in which cellular pH was determined and for all nuclear isolation and immunofluorescence experiments. Cells were incubated at 37°C in a 95% air-5% CO 2 incubator. Control monolayers were incubated in the presence of 10 mM dextrose without metabolic inhibitors. In experiments designed to study the effect of ATP depletion followed by recovery of ATP levels, cells were exposed to cyanide and 2-deoxyglucose for 1 h as described, after which buffer was replaced with fresh KHB buffer containing 10 mM dextrose. This maneuver partially restores cellular ATP levels (36). In some experiments, cells were exposed to 5 mM cyanide in the presence of 10 mM dextrose. This has been shown to result in a negligible reduction of cellular ATP in MDCK cells (2).
Extraction and Partial Purification of cPLA 2 and Separation of Insoluble and Soluble Fractions-At the end of the period of exposure to chemical anoxia or to A23187, cells from each group were washed twice with 1ϫ PBS with 2 mM EGTA. Cells were harvested into 0.5 ml of homogenization buffer containing 2 mM EGTA, 120 mM NaCl, 50 mM Tris, pH 8.0, 50 mM ␤-glycerophosphate, 1 mM sodium vanadate, 10 M PMSF, 10 M leupeptin. Cells were sonicated at 4°C at 40-watt output and 40% duty cycle. Proteins were measured by the Coomassie dye method (Bio-Rad, Richmond, VA). The resulting homogenates were adjusted for protein and centrifuged for 1 h at 100,000 ϫ g at 4°C. The supernatant represents the soluble fraction. Pellets were washed twice with 1 ϫ PBS/2 mM EGTA and resuspended in homogenization buffer with 0.01% Triton X-100. In certain experiments, in which cPLA 2 activity was measured in the insoluble or pellet fraction, pellets were resuspended in homogenization buffer, which contained 1 M KCl and centrifuged a second time to optimize dissociation of cPLA 2 from the membrane fraction (29). In certain experiments pooled supernatants were applied to a Mono Q column (Amersham Pharmacia Biotech, Sweden), which had been equilibrated with 50 mM Tris-HCl (pH 7.4) containing 1 mM EDTA. After washing the column with the same buffer, PLA 2 was eluted with a 50-ml linear NaCl gradient of 0.12 to 1.0 M in 50 mM Tris-HCl (pH 7.4) containing 1 mM EDTA. 1-ml fractions were collected and assayed for PLA 2 activity.
Nuclear Isolation-Cells were washed with 1 ϫ PBS/2 mM EGTA, harvested by scraping and Dounce-homogenized in a buffer containing 5 mM Tris, pH 7.4, 5 mM KCl, 1.5 mM MgCl 2 , 2 mM EGTA, 1 mM DTT, 0.2 mM PMSF, 10 M leupeptin. Lysate protein content was determined, and the homogenate was centrifuged at 500 ϫ g at 4°C. The supernatant was removed, and the nuclear fraction was resuspended in buffer containing 10 mM NaCl, 10 mM Tris, pH 7.4, 5 mM EDTA, 1% Triton X-100, 1 mM PMSF, and 10 M leupeptin. The supernatant is referred to as the cytosolic fraction. Equal amounts of protein from total cell lysates, nuclear and cytosolic fractions were separated by SDS-PAGE.
Phospholipase Samples were vortex-mixed for 2 min and centrifuged at 20,800 ϫ g for 2 min. 150 l of the organic phase was transferred to an Eppendorf microcentrifuge containing 800 l of heptane and 100 mg of silica gel, which was again vortex-mixed and centrifuged. 800 l of supernatant was removed, and radioactivity was measured by liquid scintillation counting.
Immunoblotting-Proteins were separated by 10% SDS-PAGE and electrophoretically transferred to Immobilon membranes (Millipore, MA). Membranes were blocked with 0.5% milk. Primary antibodies included a polyclonal antibody raised in rabbits against an aminoterminal fragment of cPLA 2 , which included the CaLB domain and which was used at a dilution of 1:2000, and anti-GFP antibody (Santa Cruz Biotechnology, CA) at a dilution of 1:200. Secondary antibody was horseradish peroxidase-conjugated anti-rabbit antibody. Membranes were developed using chemiluminescence (Amersham Pharmacia Biotech).
Immunofluorescence-LLC-PK 1 were transfected with pEGFP or pEGFP-cPLA 2 by DEAE dextran. Cells were plated onto coverslips 24 h prior to transfection. 200 l of 1ϫ phosphate-buffered saline (PBS) containing DEAE-dextran (10 mg/ml), and chloroquine (2.5 mM) was added to 5 ml of DMEM containing 10% NuSerum (Collaborative Research, Bedford, MA). DNA (40 ng/ml) was added, and the chloroquine/ DEAE dextran/DNA mixture was layered onto cells (1 ml/well). After a 4-h incubation at 37°C, the chloroquine/DEAE dextran/DNA mixture was removed and cells were exposed to 10% Me 2 SO at room temperature for exactly 2 min. Cells were washed with 1 ϫ PBS, and fresh DMEM containing 10% fetal calf serum was added. Cells were fixed in 4% paraformaldehyde with 0.1% Triton.

Determination of [Ca 2ϩ ] of Lysates and Membrane Fractions-Lysate [Ca 2ϩ
] was determined using the fluorescent probe, Calcium Green (Molecular Probes, OR) (39). Lysates of control and ATP-depleted cells were incubated with Calcium Green salt. Fluorescence was excited at 420 and 488 nm with a xenon arc lamp (Photon Technologies International, Inc. (PTI), Lawrenceville, NJ). Light output was measured at 531 nm with a photon-counting photomultiplier tube device detection system. [Ca 2ϩ ] was calibrated with a 1 mM [Ca 2ϩ ] solution (R max ) and a 100 M EGTA solution (R min ). Data analysis was performed using PTI software.
Determination of pH i -pH i was determined using the fluorescent probe BCECF (Molecular Probes, OR) (40,41). Cells grown on coverslips were incubated with 10 M BCECF-AM in HEPES buffer for 20 min at 37°C. Cells were washed to removed extracellular BCECF-AM, and coverslips were transferred to a water-jacketed cuvette maintained at 37°C. Cells were alternatively excited at 490 and 440 nm, and emission was measured at 535 nm. pH was calibrated using nigericin (40). After pH measurement, control MDCK cells were washed and incubated in buffer containing (in mM) 20 NaCl, 130 KCl, 1 MgCL 2 , 1 CaCl 2 , 20 HEPES, pH 7.5, and 10 M nigericin for 20 min. This buffer was replaced with fresh nigericin-containing buffer of pH 6.5, and fluorescence was measured after 20-min equilibration. Fluorescence measurements of six experiments were averaged to provide the calibration standard.
Statistics-All data are expressed as mean Ϯ S.E. Student's t test was used for the comparison of two groups of data. Analysis of variance (ANOVA) was used when more than two groups were compared. Following the ANOVA, the Bonferroni correction was used. p Ͻ 0.01, obtained using Student's t test, was considered significant.

RESULTS
Characterization of PLA 2 Activity-PLA 2 activity eluted as a single fraction on Mono Q anion exchange chromatography of MDCK cell supernatants (Fig. 1a). To exclude activity from co-eluting group I or group II PLA 2 isoforms, we tested peak fractions for inhibition by dithiothreitol (DTT) using 1-stearoyl-2-[1-14 C]arachidonyl-phosphatidylethanolamine as substrate. There was no inhibition of activity of partially purified PLA 2 from MDCK cells by DTT exposure for 30 min. By contrast, incubation of purified platelet Group II PLA 2 with 5 mM DTT markedly inhibited enzyme activity (Fig. 1b).
PLA 2 Activity in Cell Fractions after ATP Depletion-Cyanide and 2-deoxyglucose resulted in a decrease in total cell lysate PLA 2 activity to 45.5% of control (Fig. 2a). Western blot analysis using an anti-cPLA 2 antibody showed no obvious decrease in total cPLA 2 mass (Fig. 2c). Most of the enzyme appeared as the faster migrating form in lysates of ATP-depleted compared with control cells, suggesting dephosphorylation of the enzyme (Fig. 2c). Exposure of cells to cyanide and 2-deoxyglucose resulted in elution of cytosolic PLA 2 activity off the Mono Q column at a lower NaCl concentration (Fig. 2b), possibly because of dephosphorylation of the enzyme (37). Because cPLA 2 activity is up-regulated by phosphorylation (14 -18), these data suggest that dephosphorylation of cPLA 2 , due to ATP depletion, decreases intrinsic enzyme activity.
To determine whether the observed effect of cyanide and deoxyglucose on PLA 2 activity is due to ATP depletion or to a direct toxic effect of cyanide, we exposed cells to cyanide in the presence of 10 mM dextrose. Exposure of MDCK cells to cyanide in the presence of dextrose has been shown to reduce cellular ATP to 40 -45% of control. This degree of ATP depletion does not result in cell injury in MDCK cells (2). There is no change in soluble or membrane PLA 2 activity in cells exposed to cyanide in the presence of dextrose (20 Ϯ 6 pmol/mg/min) compared with control (18 Ϯ 4 pmol/mg/min)(data not shown).
Cell lysates were centrifuged at 100,000 ϫ g, resulting in a cytosolic fraction and an insoluble pellet that includes total cell membrane. Soluble, or S100, PLA 2 activity in untreated cells was 18 Ϯ 3 pmol/mg of protein/min. After 15 min of exposure to cyanide and deoxyglucose, S100 PLA 2 activity was decreased to 62 Ϯ 9% of control with a further decrease to 36 Ϯ 7% by 120 min of exposure (p Ͻ 0.01%) (Fig. 3a). This decrease in S100 PLA 2 activity, enriched in cPLA 2 , in ATP-depleted cells was greater than the decrease observed in total cell lysates.
To determine whether this observed decrease in soluble cPLA 2 activity was partially due to translocation of the enzyme to membrane or entirely to dephosphorylation of the enzyme, we measured activity in the insoluble fraction of cell lysates. PLA 2 activity in the 100,000 ϫ g pellet, representing the membrane fraction, was considerably lower than that measured in supernatants under all conditions studied. In membranes from control cells PLA 2 activity was 3 Ϯ 0.6 pmol/mg/min. The markedly low membrane activity compared with soluble PLA 2 activity may be partly due to dilution of labeled substrate by the excess of unlabeled phospholipid in the membrane as suggested by Channon (28), which would result in an apparent decrease in overall activity. There was an increase in activity to 228 Ϯ 33% of control in the insoluble fractions of cells after 15 min of exposure to cyanide and deoxyglucose (p Ͻ 0.01) (Fig.  3b). There was no further increase at 30, 60, 120, or 240 min. These data suggested that the striking decrease in soluble cPLA 2 activity is at least partly due to a shift in enzyme activity from the soluble to insoluble fraction of cell lysates.
To determine whether a shift in enzyme mass from soluble to insoluble fractions occurred, we performed immunoblot analy- , and an increase in electrophoretic mobility. a, lysates of ATP-depleted and control MDCK cells were tested for in vitro PLA 2 activity. There was a marked reduction in PLA 2 activity in the lysates of ATP-depleted cells compared with control cell lysates. b, S100 fractions obtained from control and ATP-depleted MDCK cells were applied to a Mono Q anion-exchange column and eluted with an NaCl gradient as in Fig. 1. PLA 2 activity in fractions of ATP-depleted cells eluted with lower [NaCl]. c, Western blot analysis of ATP-depleted MDCK cell lysates, using an anti-cPLA 2 antibody, shows no decrease in mass of cPLA 2 compared with lysates from control cells. An increase in electrophoretic mobility was observed with ATP depletion (lanes 3 and 4), possibly due to dephosphorylation of the enzyme. Control cell lysates were run in lanes 1, 2, 5, and 6.
FIG. 1. PLA 2 activity from partially purified S100 fraction is not inhibited by DTT. a, pooled control MDCK cell S100 fractions were applied to a Mono Q anion-exchange column equilibrated with 50 mM Tris, pH 7.4, buffer. PLA 2 activity was eluted with a 50-ml linear NaCl gradient of 0.12 to 1.0 M. PLA 2 activity eluted as a single fraction. b, peak fractions eluted from the Mono Q column were incubated with DTT and tested for in vitro PLA 2 activity. DTT had no effect on MDCK S100 PLA 2 activity but inhibited activity of purified Group II platelet PLA 2 which was used as a control. sis of soluble and insoluble fractions from control cells and from ATP-depleted cells. Western blot analysis of soluble and insoluble fractions is shown in Fig. 4a. There is a decrease in cPLA 2 protein in the S100 fraction of ATP-depleted cells when compared with control cells. In contrast, there is an increase in total cPLA 2 mass in the insoluble fractions of ATP-depleted cells compared with insoluble fractions from control cells. The shift in cPLA 2 enzyme mass from soluble to insoluble fraction is not reversed by the addition of exogenous ATP to cell lysates. Thus cPLA 2 enzyme mass shifts from soluble to insoluble fractions of lysates of ATP-depleted cells.
To determine if translocation resulting from ATP depletion is reversible, we exposed cells to cyanide and deoxyglucose for 2 h, then removed the metabolic inhibitors and added 10 mM dextrose. Although 2 h of exposure to cyanide and deoxyglucose decreased soluble PLA 2 activity to 27 Ϯ 5% of control (p Ͻ .01), subsequent exposure to dextrose for 1 h, after metabolic inhibitors were removed, restored PLA 2 activity to 98 Ϯ 12% of control (not significant compared with control) (Fig. 4b).
Others (25,26) have shown that, although transient increases in [Ca 2ϩ ] i target cPLA 2 to the nuclear membrane in vivo, the association of cPLA 2 with membranes is determined by the final [Ca 2ϩ ] of cell lysate (29,42). Clearly this is not the case after ATP depletion, because lysate [Ca 2ϩ ] levels from ATP-depleted and control cells were both determined to be less than 10 nM (Fig. 5). Because the buffer in which soluble and insoluble fractions are obtained has a very low [Ca 2ϩ ], these data indicate that ATP depletion results in the association of cPLA 2 with the insoluble fractions of cells, which cannot be reversed by reducing ambient [Ca 2ϩ ].
Translocation of cPLA 2 to Nuclei of ATP-depleted Cells-It is known that cPLA 2 translocates to nuclear membranes upon activation of cells with some agonists (25,26). To determine whether cPLA 2 translocated to nuclei of ATP-depleted cells, MDCK cells were harvested into 2 mM EGTA buffer and nuclear and cytosolic fractions were separated by 500 ϫ g centrifugation. Whereas cPLA 2 is demonstrated predominantly in the cytosol of control cells, the enzyme is found almost exclusively in the nuclear fraction of ATP-depleted cells (Fig. 6a).
To confirm this result, LLC-PK 1 cells were transfected with pEGFP-C1-cPLA 2 , which drives expression of cPLA 2 fused to green fluorescent protein (GFP) at the amino terminus. To determine whether the GFP tag interfered with [Ca 2ϩ ]-regulated translocation, we added Ca 2ϩ to lysates of transfected (non-ATP-depleted) cells prior to isolating nuclei. Whereas cPLA 2 is present predominantly in the cytosolic fractions from [Ca 2ϩ ]-free lysates, cPLA 2 is present in the nuclear fraction of lysates to which calcium was added, demonstrating that GFPtagged cPLA 2 translocates in response to [Ca 2ϩ ] (Fig. 6b). Immunofluorescence microscopy of control and ATP-depleted GFP-cPLA 2 -expressing LLC-PK 1 cells revealed that GFP-cPLA 2 is present in a homogenous distribution in the cytosol of untreated LLC-PK 1 cells. By contrast, GFP-cPLA 2 is present in a perinuclear distribution in ATP-depleted cells (Fig. 6c).
BAPTA Blocks cPLA 2 Nuclear Translocation-Thus, cPLA 2 translocates to the nuclear membrane of ATP-depleted cells FIG. 3. ATP depletion decreases PLA 2 activity in the S100 fraction and increases activity in the insoluble fraction. a, PLA 2 activity was measured in the S100 fraction of ATP-depleted and control cells. PLA 2 activity was decreased to 62 Ϯ 9%, 62 Ϯ 7%, 48 Ϯ 9%, 36 Ϯ 7%, and 23 Ϯ 7% of control cell fractions at 15, 30, 60, 120, and 240 min of exposure to cyanide and 2-deoxyglucose in the absence of metabolic substrate. *, p Ͻ 0.01% compared with control. b, PLA 2 activity was measured in the membrane fraction of cells exposed to 15, 30, and 60 min of cyanide and 2-deoxyglucose. There were increases in membraneassociated PLA 2 activity to 228 Ϯ 33%, 208 Ϯ 42%, and 206 Ϯ 32% of control at 15, 30, and 60 min. *, p Ͻ 0.01% compared with control. FIG. 4. ATP depletion causes shift of cPLA 2 mass from S100 to insoluble fractions. The ATP depletion-induced decrease in S100 PLA 2 activity is reversed by addition of metabolic substrate. a, proteins of S100 and insoluble fractions of ATP-depleted and control cells were separated by SDS-PAGE and blotted with an anti-cPLA 2 antibody. There is a decrease in cPLA 2 mass in the S100 fractions of ATP-depleted cells compared with control cells, whereas there is an increase in cPLA 2 mass in the insoluble fractions of ATP-depleted cells when compared with control cells . b, PLA 2 activity was measured in the S100 fraction of cells exposed to 2 h of cyanide (cn) and 2-deoxyglucose (dog) followed by 1 h of exposure to 10 mM dextrose (dex) after metabolic inhibitors were removed. S100 PLA 2 activity was compared with that of cells treated with cyanide and 2-deoxyglucose for 2 h as well as to dextrose-treated control cells. After 2 h of cyanide and 2-deoxyglucose, PLA 2 activity was reduced to 27 Ϯ 5% of control. After 2 h of cyanide and 2-deoxyglucose followed by 1 h of dextrose, PLA 2 activity was 98 Ϯ 12% of control. *, p Ͻ 0.01 compared with control. and remains associated with nuclear fractions despite homogenization of cells into EGTA-containing buffer. ATP depletion of MDCK cells has been reported to result in an increase in [Ca 2ϩ ] i from 112 to 649 nM within minutes (43), which is sufficient to induce cPLA 2 translocation to insoluble fractions (7). To determine whether a transient increase in [Ca 2ϩ ] i plays a role in the ATP depletion-induced translocation of cPLA 2 , cells were depleted of ATP depletion after preincubation with BAPTA-AM, which upon entry into cells is cleaved to BAPTA, which serves to chelate Ca 2ϩ and prevent the increase in [Ca 2ϩ ] i . cPLA 2 -transfected LLC-PK 1 cells were preincubated with 100 M BAPTA or with Me 2 SO in HEPES buffer in the absence of Ca 2ϩ or Mg 2ϩ . Cyanide and 2-deoxyglucose or dextrose were added to cells for 2 h. BAPTA, but not Me 2 SO, prevented the ATP depletion-induced translocation of cPLA 2 to nuclear membrane (Fig. 7a). MDCK cells were similarly preincubated with BAPTA or Me 2 SO. BAPTA, but not Me 2 SO, partially inhibited the association of cPLA 2 with nuclear fractions of ATP-depleted cells (Fig. 7b).

A23187-induced Nuclear Translocation of cPLA 2 Is Reversed When Homogenate [Ca 2ϩ ] Is Reduced-The effect of BAPTA on ATP depletion-induced translocation and association of cPLA 2 with nuclear fractions suggests that cPLA 2 translocation is dependent upon a transient increase in [Ca 2ϩ ] i although not reversed upon lowering of final homogenate [Ca 2ϩ ] with EGTA.
To demonstrate that there is a fundamental difference in the character of cPLA 2 binding to membranes when cells are ATPdepleted, control cells were treated with 2 M A23187 for 20 -120 min (Fig. 8, a and b). GFP-cPLA 2 localizes to a perinuclear region of A23187-treated, transfected LLC-PK 1 (Fig. 8a, right  panel) but is localized diffusely in the cytosol of transfected vehicle-treated cells (left panel). After these cells are homogenized in low [Ca 2ϩ ] buffer (Ͻ10 nM) cPLA 2 dissociates from the nuclear fraction. cPLA 2 associates predominantly with the cytosolic fraction of A23187-treated MDCK cells (Fig. 8b). Thus, although transient increases in [Ca 2ϩ ] i cause cPLA 2 to localize to the nuclear membrane in vivo, in contrast to ATP-depleted cells, the association of cPLA 2 with nuclear fractions of ATPreplete cells is disrupted by reduction of lysate [Ca 2ϩ ].
ATP depletion also causes a decrease in pH i (44). To determine whether a decrease in pH i might account for the association of cPLA 2 with nuclear fractions, we harvested ATP-replete MDCK cells in EGTA-containing buffers varying in pH from 4.5 to 7.5. Although there is limited cPLA 2 association with the nuclear fraction of cells harvested into lysis buffer at pH of 4.5 or below, cPLA 2 associates exclusively with the cytosolic fraction of cells harvested into buffer with a pH of 6.0 or greater (Fig. 9a). Cell pH i , estimated by the intracellular fluorescent probe, BCECF, decreases after 2 h of ATP depletion but remains above 6.5 (Fig. 9b). Thus, a decrease in pH does not explain the ATP depletion-induced translocation of cPLA 2 .
Okadaic Acid Partially Prevents the ATP Depletion-induced Nuclear Translocation of cPLA 2 -ATP depletion causes the dephosphorylation of cellular proteins (45). To determine whether the inhibition of dephosphorylation of either cPLA 2 or of another cellular protein might affect cPLA 2 association with nuclear fractions in response to ATP depletion, we preincubated cells in phosphatase 2A inhibitor, okadaic acid, prior to exposure to cyanide and 2-deoxyglucose. Okadaic acid, but not  6. cPLA 2 translocates to nuclei of ATP-depleted cells. a, control and ATP-depleted MDCK cells were harvested into EGTAcontaining buffer and Dounce-homogenized. Nuclei were isolated by 500 ϫ g centrifugation. cPLA 2 is present predominantly in the cytosolic fraction of control cell lysates and the nuclear fraction of ATP-depleted cell lysates. cPLA 2 is present in comparable amounts in the lysates of both groups. b, LLC-PK 1 cells were stably transfected with pEGFP-C1-cPLA 2 . Non-ATP-depleted cells were harvested into EGTA-containing buffer and Dounce-homogenized. Nuclear and cytosolic fractions were then prepared from this lysate or lysate to which 10 mM Ca 2ϩ was added. GFP-cPLA 2 is present almost exclusively in the cytosol fraction of [Ca 2ϩ ]-free lysates and in the nuclear fraction of [Ca 2ϩ ]-containing lysates. c, pEGFP-C1-cPLA 2 -transfected LLC-PK 1 cells were plated on coverslips and treated with either dextrose (control), or cyanide and 2-deoxyglucose (ATP depletion). Cells were fixed and exposed to anti-GFP antibody followed by Cy3-conjugated anti-rabbit antibody. GFP-cPLA 2 appears diffusely in the cytosol of control cells, whereas the fusion protein localizes to the perinuclear region of ATP-depleted cells. the Me 2 SO vehicle, partially inhibits the nuclear translocation of cPLA 2 (Fig. 10). DISCUSSION We have demonstrated that exposure of MDCK cells to cyanide and deoxyglucose results in a marked decrease in total lysate and soluble PLA 2 activity. This decrease is seen as early as 15 min of addition of cyanide/2-deoxyglucose with further decreases in activity at 2 and 4 h. The decrease in PLA 2 activity is likely due to ATP depletion rather than to a direct toxic effect of cyanide, because cyanide in the presence of dextrose, which causes a less marked decrease in cellular ATP, has no effect on PLA 2 activity.
The decrease in PLA 2 activity induced by cyanide and deoxyglucose is rapidly reversed by the addition of dextrose, presum-ably by restoration of cellular ATP levels to a critical threshold. Thus, ATP depletion induces a rapidly reversible decrease in PLA 2 activity. Partial purification of MDCK cell supernatants on a Mono Q column suggests that the predominant [Ca 2ϩ ]-dependent PLA 2 isoform active against a diacylphospholipid substrate is cPLA 2 .
Two mechanisms may explain the decrease in soluble cPLA 2 activity that is induced by ATP depletion. The first possibility is a reduction in the phosphorylation state of the enzyme due to a reduction in the energy charge ratio in the cell that results in a decrease in the intrinsic activity of the enzyme. The immunoblot analysis of total cell lysates suggests that cPLA 2 is dephosphorylated after ATP depletion, resulting in increased electrophoretic mobility of cPLA 2 . Dephosphorylation may also explain elution of cPLA 2 protein at a lower NaCl concentration, which is likely due to a decreased affinity to the anion resin (37). These data are not surprising. ATP, which is necessary to maintain phosphorylation of the enzyme, is reduced to less than 5% of control levels in this model of anoxia (2). Kobryn et al. (45) have demonstrated a decrease in protein phosphorylation in suspended rabbit proximal tubules under anoxic conditions. Numerous studies have shown that phosphorylation of cPLA 2 modulates its activity (14 -17, 19, 22, 23). Our data suggests that cPLA 2 is partially phosphorylated under control conditions in MDCK cells grown in 10% serum. ATP depletion results in dephosphorylation and a decrease in intrinsic enzyme activity.
We also demonstrate translocation of an isoform of cPLA 2 from cytosol to the nuclear membrane of ATP-depleted cells. The nuclear translocation of cPLA 2 likely explains the marked decrease in S100 activity and the increase in membrane activity that was observed after 15 min of anoxia.
Other studies have suggested translocation of a PLA 2 isoform after ischemia or ATP depletion. In a rat model of ischemia, our laboratory has demonstrated a decrease in PLA 2 activity in cytosolic fractions and an increase in activity in mitochondrial fractions after 45 min of ischemia (4). Portilla et al. (5) demonstrated a shift in both PLA 2 activity and in mass of a 40-kDa isoform of PLA 2 to membrane fractions of rabbit proximal tubules after anoxia. PKC␥ has also been shown to translocate to cell membranes in response to ischemia in the rat model (46). This is particularly interesting given the regions of homology shared by cPLA 2 and PKC␥.
The mechanism underlying translocation of cPLA 2 to the nuclear membrane of ATP-depleted cells is unknown. ATP depletion causes both an increase in [Ca 2ϩ ] i (43) and a decrease FIG. 8. A23187 causes nuclear translocation of cPLA 2 , which is reversed when cells are homogenized in low-[Ca 2؉ ] buffer. a, non-ATP-depleted pEGFP-cPLA 2 -transfected LLC-PK 1 cells were treated with 2 M A23187 for 30 min and examined by immunofluorescence microscopy. GFP-cPLA 2 localizes to a perinuclear region of A23187-treated cells, whereas it appears diffusely in the cytosol of vehicle-treated control cells. b, non-ATP-depleted MDCK cells were treated with A23187 for 30 -120 min and harvested into EGTA-containing buffer. cPLA 2 associates predominantly with the cytosolic fraction of both control and A23187-treated cells.
FIG. 9. The ATP depletion-induced decrease in pH is not sufficient to cause nuclear translocation of cPLA 2 . a, ATP-replete MDCK cells were harvested into buffers of pH between 4.0 to 7.5, and nuclear and cytosolic fractions were isolated. cPLA 2 associates almost exclusively with the cytosolic fractions (c) of lysates at pH 6.0 -7.5. cPLA 2 is detected in the nuclear fractions (n) of lysates when pH is equal or less than 4.5. b, MDCK cells grown on coverslips were loaded with BCECF-AM and excited at 440 and 490 nm. Emission was detected at 535 nm, and values are expressed as the ratio of counts obtained at 490/440. Cellular pH decreases after 2 h of ATP depletion compared with control but remains above 6.5 in all conditions. in pH i (44). We were able to partially inhibit the translocation of cPLA 2 by blocking the rise in [Ca 2ϩ ] i with BAPTA, which suggests that an ATP depletion-induced [Ca 2ϩ ] i increase contributes to cPLA 2 translocation. There is a fundamental difference, however, between the nuclear translocation that occurs with ATP depletion when compared with that which occurs by increasing [Ca 2ϩ ] in ATP-replete cells. Although A23187 targets cPLA 2 to the nuclear membrane of ATP replete cells, this association of cPLA 2 with the nuclear fractions is lost once the cells are homogenized in a buffer with a [Ca 2ϩ ] of Ͻ10 nM. Thus, a transient increase in [Ca 2ϩ ] i is necessary for nuclear translocation, but it is insufficient to cause the persistent association of cPLA 2 with nuclear fractions of cells.
A decrease in pH that occurs with ATP depletion cannot explain nuclear translocation of cPLA 2 . The decrease in cellular pH with ATP depletion of MDCK cells is very modest compared with that necessary to cause nuclear translocation.
ATP depletion decreases the phosphorylation of cellular proteins (45). Okadaic acid inhibits serine/threonine-specific phosphatases, and in particular (though not exclusively) protein phosphatase 2A (47). Our data indicate that inhibition of dephosphorylation by okadaic acid partially inhibits the ATP depletion-induced translocation of cPLA 2 . Although we do not know whether okadaic acid alters the phosphorylation state of cPLA 2 or of other proteins that may interact with cPLA 2 , these data suggest that phosphorylation/dephosphorylation is an additional regulatory component in [Ca 2ϩ ]-dependent nuclear membrane association of cPLA 2 .
Thus, ATP depletion causes translocation of cPLA 2 to the nucleus of MDCK cells. An increase in [Ca 2ϩ ] i likely contributes to the translocation but cannot completely explain the ATP depletion-induced persistent translocation of cPLA 2 despite lowering ambient [Ca 2ϩ ]. The partial abrogation of ATP depletion-induced cPLA 2 translocation by the phosphatase inhibitor okadaic acid suggests a role for phosphorylation/dephosphorylation of either cPLA 2 or other proteins in cPLA 2 translocation. These novel observations suggest that there are mechanisms in addition to changes in [Ca 2ϩ ] that modulate cPLA 2 translocation. cPLA 2 plays an important role in physiologic (8,9,48) and pathophysiologic states (10,11,49) and has recently been shown to interact with nuclear proteins, suggesting an intranuclear role (50). Because cellular ATP depletion is an important component of pathophysiologic states that result in end organ ischemia (8,51), the modulation of cPLA 2 trafficking by ATP depletion may contribute to ischemic pathophysiology. A better understanding of the regulation of cPLA 2 translocation under control and ATP-depleted conditions may allow for the development of therapeutic strategies in these conditions.